Re: [mmox] Creating walled gardens considered harmful

[Morgaine <[e-mail]>]

On Sat, Mar 28, 2009 at 6:16 AM, Jon Watte <[e-mail]> wrote:

> Jon wrote:
>
>   There is no actual interoperability achieved if those are
>   standardized. /For a user of any other virtual world platform than
>   SecondLife/OpenSim, these services would give them zero value./
>
>
> Morgaine wrote:
>
>> I don't understand your last sentence at all, and therefore the remainder
>> of your post was lost on me.  The above 3 points apply identically to every
>> virtual world under the sun, and you agreed to that previously <
>> http://www.ietf.org/mail-archive/web/mmox/current/msg01216.html>.
>>
>
> Yes, they are necessary, but not sufficient, to provide any kind of useful
> interop. I thought I made that pretty clear, but I assume in good faith that
> you're somehow not seeing it.



Of course they're not sufficient!  They're not meant to be.  We're in the
process of breaking down the problem space, and each component of that space
is in itself not sufficient.  Putting all the required components together
later will yield a sufficient solution.  That's the engineering process in
action.

This is why continually retreating back to the monolithic corner positions
is a disaster for our work --- it derails all the engineering progress we
have achieved so far in problem decomposition.  We *know* those positions
already, from pre-BoF, they don't need to be restated.

Having sorted out *teleport* and shown it not to be an area of
incompatibility, we can now proceed to the next area of difficulty.  Moving
one step at a time we can solve this problem.  Just proclaiming that each
small step does not solve the end goal achieves nothing and sets back the
analysis.

Instead, please make sure that all the component parts required by your use
cases are in our problem space, and then we can handle the issues one at a
time.


Morgaine.








On Sat, Mar 28, 2009 at 6:16 AM, Jon Watte <[e-mail]> wrote:

> Jon wrote:
>
>   There is no actual interoperability achieved if those are
>   standardized. /For a user of any other virtual world platform than
>   SecondLife/OpenSim, these services would give them zero value./
>
>
> Morgaine wrote:
>
>> I don't understand your last sentence at all, and therefore the remainder
>> of your post was lost on me.  The above 3 points apply identically to every
>> virtual world under the sun, and you agreed to that previously <
>> http://www.ietf.org/mail-archive/web/mmox/current/msg01216.html>.
>>
>
> Yes, they are necessary, but not sufficient, to provide any kind of useful
> interop. I thought I made that pretty clear, but I assume in good faith that
> you're somehow not seeing it. Let's try again. The logic is simple
> economics:
>
> Because they are not sufficient in themselves, there is no incentive for
> vendors who don't have all the other pieces also in place, to implement
> those three, because it's costly work with no actual benefit. (Not even Open
> Source engineering is free, because there is always opportunity cost for the
> project in question.)
>
> Thus, even if those things were standardized, there would be no incentive
> for Multiverse.net to implement them, or for Proton Media, or for Qwaq, or
> for Habbo Hotel, or for Forterra. There would only be incentive to implement
> them for Open Sim and Second Life, because they already have enough
> interoperability (thanks to the goals of Open Sim itself). Think about it --
> if those three were standardized, then the proposal to most virtual world
> vendors in the world would be:
> "Here are three things you can implement. Once you have implemented them,
> there is no additional visible capability for your platform or your users,
> unless you also implement a bunch of other, un-standardized things that have
> to do with simulation and state replication."
> If you had a virtual world platform, and you could direct your engineers to
> spend time on that, or spend time on something that actually delivered a
> feature that your end users would see and appreciate, which would you
> choose?
>
> So, if there was a standard that standardized those three things, but
> nothing else, then the only platforms that would adopt those standards would
> be the ones that already proposed OGP out of the Second Life AWG, because
> they already have interop in the other parts. The IETF standardization
> effort would then be pretty much a big no-op (compared to staying in the
> Second Life specific AWG), assuming that the goal of going to the IETF to
> provide standardization that is adopted by a broader group.
>
> It may be that I'm missing something. It may be that the users of Areae, or
> CokeWorld, or Avatar Reality would actually gain something tangible from
> those worlds implementing the three services in question, and I just haven't
> been able to see the use case that would be enabled. If that is the case,
> perhaps you could help me by describing a detailed use case of what is
> actually being enabled by standardizing only those three services.
>
> I think that is about as clearly and to the point I can describe the
> problem.
>
> Let me know if there is something else I can do to put this point in a
> different light, because I think it is important, and speaks to the question
> about whether we even should try to do something the IETF way. If we're not
> willing to sit down and hammer out something that, if implemented, would
> provide value to the users of most virtual world platforms without relying
> on large amounts of other, out-of-band engineering to go with it, then
> whatever we do will be in vain.
>
> Sincerely,
>
> jw
>
>